Systems genetics of obesity in an F2 pig model by genome-wide association, genetic network, and pathway analyses

ORIGINAL RESEARCH ARTICLEpublished: 09 July 2014

doi: 10.3389/fgene.2014.00214

Systems genetics of obesity in an F2 pig model bygenome-wide association, genetic network, and pathwayanalysesLisette J. A. Kogelman , Sameer D. Pant , Merete Fredholm and Haja N. Kadarmideen*

Animal Genetics, Bioinformatics and Breeding Section, Department of Veterinary Clinical and Animal Sciences, Faculty of Health and Medical Sciences, Universityof Copenhagen, Copenhagen, Denmark

Edited by:

Natalia Polouliakh, Sony ComputerScience Laboratories Inc., Japan

Reviewed by:

Olga Papadodima, National HellenicResearch Foundation, GreeceSachiyo Aburatani, National Instituteof Advanced Industrial Science andTechnology, Japan

*Correspondence:

Haja N. Kadarmideen, AnimalGenetics, Bioinformatics andBreeding Section, Department ofVeterinary Clinical and AnimalSciences, Faculty of Health andMedical Sciences, University ofCopenhagen, Grønnegårdsvej 7,1870 Frederiksberg C, Copenhagen,Denmarke-mail: [email protected]

Obesity is a complex condition with world-wide exponentially rising prevalence rates,linked with severe diseases like Type 2 Diabetes. Economic and welfare consequenceshave led to a raised interest in a better understanding of the biological and geneticbackground. To date, whole genome investigations focusing on single genetic variantshave achieved limited success, and the importance of including genetic interactions isbecoming evident. Here, the aim was to perform an integrative genomic analysis in an F2pig resource population that was constructed with an aim to maximize genetic variationof obesity-related phenotypes and genotyped using the 60K SNP chip. Firstly, GenomeWide Association (GWA) analysis was performed on the Obesity Index to locate candidategenomic regions that were further validated using combined Linkage DisequilibriumLinkage Analysis and investigated by evaluation of haplotype blocks. We built WeightedInteraction SNP Hub (WISH) and differentially wired (DW) networks using genotypiccorrelations amongst obesity-associated SNPs resulting from GWA analysis. GWA resultsand SNP modules detected by WISH and DW analyses were further investigated byfunctional enrichment analyses. The functional annotation of SNPs revealed severalgenes associated with obesity, e.g., NPC2 and OR4D10. Moreover, gene enrichmentanalyses identified several significantly associated pathways, over and above the GWAstudy results, that may influence obesity and obesity related diseases, e.g., metabolicprocesses. WISH networks based on genotypic correlations allowed further identificationof various gene ontology terms and pathways related to obesity and related traits, whichwere not identified by the GWA study. In conclusion, this is the first study to developa (genetic) obesity index and employ systems genetics in a porcine model to provideimportant insights into the complex genetic architecture associated with obesity and manybiological pathways that underlie it.

Keywords: obesity index, animal model, high-throughput genotype data, systems genetics, WISH network

INTRODUCTIONObesity, a complex condition characterized by excessive accu-mulation of body fat, is an exponentially growing public healthproblem associated with several severe diseases like Type 2Diabetes, cardiovascular diseases, and various types of cancers(Bener et al., 2005). Obesity in humans is influenced by environ-mental, epigenetic, and genetic factors. Animal (model) studiesshow that weight gain and adiposity are related to genetic dif-ferences in eating behavioral patterns (Do et al., 2013), and ithas also been shown that epigenetics play a critical role (e.g.,nutritional status during pregnancy/fetal programming events)in determining effective partitioning of nutrients to maintain-ing essential physiological functions (Hou et al., 2013; Do et al.,2014). However, these environmental and epigenetic factors aredifficult to manage, and therefore there is substantial interest ingaining more knowledge about the genetic background of obe-sity. Results from GWA studies performed on obesity-related

traits, e.g., body mass index (BMI) and waist-hip ratio (WHR),are to date unsatisfactory in unraveling the genetic backgroundfor obesity. While phenotypes like BMI are known to be highlyheritable (40–70%), the largest GWA study on BMI has onlybeen able to identify 32 candidate loci that together explainonly 1.45% of the inter-individual variation (Speliotes et al.,2010). Similarly in the case of WHR, only 14 candidate lociwere identified that together explain only 1.03% of the vari-ance in WHR compared to an estimated heritability between 22and 61% (Heid et al., 2010). Beside gene-environment interac-tions (Qi and Cho, 2008), it is becoming increasingly evidentthat genome-wide genetic interactions, which are not taken intoaccount by GWA studies, could play a key role to these discrepan-cies (Cordell, 2009). For example, the well-known obesity-relatedFTO gene interacts with APOE which in turn, is associatedwith Alzheimer’s disease (Keller et al., 2011) and with MC4R,resulting in a higher chance of breast cancer (Cunha et al.,

www.frontiersin.org July 2014 | Volume 5 | Article 214 | 1

http://www.frontiersin.org/Genetics/editorialboard



http://www.frontiersin.org/Genetics/about

http://www.frontiersin.org/Genetics

http://www.frontiersin.org/journal/10.3389/fgene.2014.00214/abstract

http://community.frontiersin.org/people/u/75274




mailto:[email protected]

http://www.frontiersin.org

http://www.frontiersin.org/Systems_Biology/archive

Kogelman et al. Systems genetics of obesity in pig model

2013). These studies demonstrate both the presence of geneticinteractions, and how these interactions could potentially influ-ence the biological relationship between obesity and severalother severe diseases. Therefore, the application of network-based genetics or systems genetics approaches directed towardinvestigating these genome-wide interactions and their functionalrelevance has the potential to provide novel biological and geneticinsights into complex traits and diseases (Kadarmideen et al.,2006; Joshi et al., 2009; Civelek and Lusis, 2014; Kadarmideen,in press).

Various systems genetics approaches have been developed tounravel the genetic background of complex diseases by distin-guishing networks, functional pathways, and underlying causalgenes (Brazhnik et al., 2002; Segal et al., 2003; Diez et al., 2010;Horvath, 2011; Kadarmideen et al., 2011; Yang et al., 2011;Kadarmideen, in press). Recently, we published the WeightedInteraction SNP Hub (WISH) network method, which identi-fies clusters of highly interconnected SNPs (modules) using highthroughput genotype data (Kogelman and Kadarmideen, 2014).Functional annotation and pathway analysis of detected modulesmay lead to the identification of biologically relevant pathways.Moreover, the differential connectivity or differential wiring ofSNPs between two subgroups of extreme phenotypes, may revealbiologically interesting candidates as they might point to differ-ences in underlying genetic regulation for manifestation of thetrait of interest.

In the present study, we applied WISH network and geneenrichment analyses in addition to GWA analysis (validated usingLDLA and haploblock methods), in order to identify individ-ual variants and pathways that may be causal for obesity usinga porcine model established for obesity studies (Kogelman et al.,2013). Animal models are helpful to study complex diseases dueto reduced costs, more controlled environment, and the possibil-ity of more controlled and extensive phenotyping in comparisonwith human studies. Pigs are a valuable model for human obesityand obesity-related diseases primarily because of comparable fea-tures that include a similar cardiovascular system, proportionallysimilar organ sizes, and comparable protein and lipid metabolism(Spurlock and Gabler, 2008). Pigs are also genetically close tohumans and therefore acknowledged as an important biomedicalmodel to study obesity (Groenen et al., 2012). The F2 pig resourcepopulation used in this study was constructed by intercrossingGöttingen Minipigs that are prone to obesity, and production pigsthat are selected for leanness over many generations. This popula-tion has a high degree of genetic variation among several obesity-and obesity related traits (Kogelman et al., 2013) and is thereforea promising resource for further analyses directed toward iden-tifying novel genetic determinants of obesity and obesity relateddiseases.

The main goal of this study was to investigate the genetic deter-mination of obesity by applying integrated systems approachesincluding GWA, LDLA, WISH network analyses of wholegenomic data and functional gene enrichment analyses, in apig model. Here we report SNPs, and associated genes, bio-logical pathways and functional ontologies identified via theapplication of these approaches to a combination of obesityphenotypes.

RESULTSOBESITY INDEXThe Obesity Index (OI) was constructed according to the selec-tion index theory (Cameron, 1997), whereby nine obesity andobesity-related (OOR) phenotypes are combined into a singlevalue representative of the genetic predisposition to obesity. Asexpected based on the experimental design, the OI followeda normal distribution within the F2 pig resource population(Figure 1), with a mean of −0.06 and a standard deviation of1.15. Accordingly, the OI can be used to select animals that aregenetically predisposed to being either extremely obese or lean.

GENOME-WIDE ASSOCIATION STUDY (GWAS)A GWAS on the OI was performed on the entire F2 pig resourcepopulation. After quality control of the high-throughput geno-type data, 40,910 markers and 538 animals were used for analysis.In total, 404 SNPs passed the Bonferroni corrected significancethreshold of P = 2.44E−8, of which 366 were assigned to theporcine assembly, all located on autosomes. The Manhattan plotrepresenting genome-wide p-values is shown in Figure 2.

In total, 289 genes were detected within 20 kb of the 366 SNPsthat passed the genome-wide significance threshold. As most ofthe significantly associated SNPs were mapped to Ensemble geneidentifiers (in total 100 unique genes), we only focused on thosegenes, and not on all the genes within a flanking distance of20 Kb. Several of those genes could be associated with obesity orobesity-related diseases. Here we focus only on arbitrarily chosenseven genes corresponding to the highest genome-wide signif-icant SNPs, detected after control of multiple testing and falsediscovery rate. Table 1 shows these SNP names, their genome-wide corrected p-values, their effects on total obesity score alongwith associated gene names.

The most significant SNP (P-value = 9.09E−17) detected bythe GWA study is located within the NPC2 gene (Nieman–Pick disease, type 2C). This gene encodes a protein that plays

FIGURE 1 | Distribution of the Obesity Index (OI), an aggregate

genotype representing the degree of obesity of all F2 animals in the F2

pig resource population.

Frontiers in Genetics | Systems Biology July 2014 | Volume 5 | Article 214 | 2

http://www.frontiersin.org/Systems_Biology




FIGURE 2 | A Manhattan plot of GWA Study single-locus P-values. The blue dash line indicates a suggestive significance threshold with adjustedBonferroni correction at Padj = 1.33E−6 and the red line indicates a highly significant threshold with adjusted Bonferroni correction at Padj = 2.67E−8.

Table 1 | Description of a selection of highly significant SNPs

associated with the Obesity Index.

SNP Chr Position Nearest gene Effect P-value

rs81396056 7 103574383 NPC2 0.63 9.09E−17

rs81238148 2 11449934 OR4D10 −0.60 1.13E−15

rs81416774 9 135204695 CACNA1E −0.76 2.56E−13

rs80910848 1 228929456 SH3GL2 −0.49 1.31E−12

rs81306707 1 177369143 CDH20 0.50 1.79E−12

rs80998394 14 29824188 AACS 0.75 1.49E−10

rs80826774 7 65275011 ADPGK 0.76 4.20E−10

a role in the regulation of cholesterol transport through thelate endosomal/lysosomal system, affecting cholesterol homeosta-sis (Storch and Xu, 2009). The second highest significant SNP(P-value = 1.13E−15) is located approximately 3 Kb downstreamfrom the OR4D10 gene, which is an olfactory receptor gene.Olfactory receptors are responsible for the perception of smell,through neuronal responses, and thereby affecting the perceptionof food flavor. Furthermore, additional highly significant SNPsmapped within the vicinity of other potentially obesity relatedgenes: e.g., CACNA1E, SH3GL2, CDH20, AACS, and ADPGK. TheCACNA1E gene encodes a protein in a voltage dependent cal-cium channel. Variants in this gene have been associated withtype 2 diabetes, insulin resistance, and impaired insulin secretionin non-diabetic subjects (Trombetta et al., 2012). The SH3GL2gene encodes Endophilin-A1, which plays a role in synaptic vesi-cle endocytosis and is associated with lipid binding (Huttnerand Schmidt, 2000). The CDH20 gene encodes a type 2 classicalcadherin, which is a calcium dependent cell-cell adhesion gly-coprotein and a prime candidate for tumor suppression (Koolset al., 2000). A study on childhood obesity in Hispanic childrenfound this gene to have a suggestive association with energy bal-ance (P-value = 5.60E−6) (Comuzzie et al., 2012). The AACSgene (acetoacetyl-CoA synthetase-like gene) mediates activationof ketone bodies for synthesis of fatty acid and cholesterol. AACSis potentially being regulated by the leptin signaling pathwayvia the brain and consequently a cause of metabolic disorders

(Narishima et al., 2009). The ADPGK gene (ADP-dependentglucokinase-like gene) regulates T-cell activation, affecting theglycolytic metabolism (glucose uptake) (Kaminski et al., 2012).All genome-wide significant GWA study results are presented inAdditional file 1.

VALIDATION OF GWAS RESULTS BY LDLA AND HAPLOTYPE BLOCKANALYSESThe GWA study revealed several highly associated SNPs, inwhich, among others, seven genes were chosen after their co-location with the highest genome-wide significant SNPs (basedon p-values; Table 1). The genomic regions around those sevengenes were further validated using the combined LDLA. We theninterrogated the LDLA validated regions using haplotype blockanalyses for genetic variants and genes that are in close vicin-ity and in high linkage disequilibrium. The LDLA approachmodeled QTLs in the middle of successive marker pairs locatedwithin the above mentioned genetic regions, and results indi-cated that QTLs modeled in all possible positions within theseseven detected genes passed the suggestive significance thresh-old (P-value = 1.6E−6) and, except for NPC2, passed the highlysignificant threshold (P-value = 3.2E−8). Regional plots of theseseven indicated genes are presented in Figure 3.

After LDLA confirmation of the seven genomic regionsdetected by GWA analysis, we further examined the validatedregions by investigating possible haplotype block structures. TheNPC2, OR4D10, and CACNA1E genes were located in small hap-lotype blocks. The NPC2 gene is located on chromosome 7. In theevaluated region (103.566–103.584 Mb) there are five small hap-lotypes, wherein the NPC2 gene is located on a small haplotypeblock of 4 Kb containing eight SNPs (Figure 4A). This haploblockdoes not contain any other genes. The OR4D10 gene is located onchromosome 2, and in the evaluated region (11441–11458 Mb)there are three small haplotype blocks, and the OR4D10 geneis located in a haplotype block of 10 Kb containing nine SNPs(Figure 4B). There are no other genes present in this haplo-type block. The CACNA1E gene is located on chromosome 9,and in the evaluated region (13596–135224 Mb) there are foursmall haplotype blocks, where CACNA1E is located in a haplotype





FIGURE 3 | LDLA regional plots for regions around seven selected

genes. The concerning gene is vertically marked with a gray bar. Theblue dash line indicates a suggestive significance threshold with

adjusted Bonferroni correction at P = 1.58E−6 and the red lineindicates a highly significant threshold with adjusted Bonferronicorrection at P = 3.17E−8.

FIGURE 4 | Haploview representation of regional haplotype block

estimates of seven selected regions, detected by GWA analysis. On thevertical line the SNPs present in the selected region, and the triangles (blacklines) represent the haplotype blocks. (A) NPC2 gene is located in a 8-SNPhaplotype block of 4 Kb size on Chromosome 7; (B) OR4D10 gene is locatedin a 9-SNP haploype block of 10 Kb size on Chromosome 2; (C) CACNA1E

gene is located in a 7-SNP haplotype block of 8 Kb size on Chromosome 9;(D) SH3GL2 gene is located in a 129-SNP haplotype block of 67 Kb size onChromosome 1; (E) CDH20 gene is located in a 16-SNP haplotype block of46 Kb size on Chromosome 1; (F) AACS gene is located in a 9-SNP haplotypeblock of 26 Kb size on Chromosome 14; (G) ADPGK gene is located in a204-SNP haplotype block of 500 Kb size on Chromosome 7.

block of 8 Kb containing seven SNPs (Figure 4C). This haplo-type block contains no other genes. The SH3GL2, CDH20, andAACS genes were located in moderately sized haplotype blocks.The SH3GL2 gene is located on chromosome 1, and in the evalu-ated region (228866–228993 Mb) there are eight small haplotype

blocks. SH3GL2 is located in the largest haplotype block (67 Kb)containing 129 SNPs (Figure 4D), containing one more gene:CNTLN (centlein, centrosomal protein). The CDH20 gene is alsolocated on chromosome 1, and in the evaluated region (177320–177418 Mb) there are three small haplotype blocks. The CDH20






gene is located on a haplotype block of 46 Kb containing 16 SNPs(Figure 4E), but does not contain any other genes. The AACS geneis located on chromosome 14, and we detected four blocks in theevaluated region (29775–29874 Mb), where the AACS gene waslocated on the haplotype block of 26 Kb containing nine SNPs(Figure 4F), but no other genes were present in this haplotypeblock. The ADPGK gene was located on chromosome 7 in a largehaplotype block of 500 Kb containing 204 SNPs (Figure 4G). Inthis haplotype block (65060–65560 Mb) there are several othergenes located: BBS4 and ARIH1 are downstream of ADPGK,TMEM202, HEXA, PARP6, PKM, and two uncharacterized pro-teins are upstream of ADPGK. According to the haplotype blockanalysis we can conclude that we can be reasonably confidentabout the genes we pinpoint, however, with the ADPGK gene wemay have a noise because of the many other genes in the samehaplotype block. Previous research has shown an average haplo-type block size of 400 Kb, of which most between 100 and 400 Kb,for several production pig lines (Veroneze et al., 2013). Since thisis an F2 population, we expected even larger haplotype block sizesthen detected here.

PATHWAY DETECTIONPathway analysis was performed using genes located within aflanking distance of 20 Kb of the genome-wide significant SNPs(P-value = 2.44E−8) identified via GWA analysis. Initially, anal-ysis using the NCBI2R R-package only identified metabolic path-ways (P-value = 0.08) that was not significant after Bonferronicorrection was applied to account for multiple comparisons.Subsequently, GeneNetwork (http://www.genenetwork.nl) wasused to detect overrepresented pathways; and phenotypes and tis-sues associated with these based on publicly available expressiondata. GeneNetwork analyses identified various cellular, trans-port related processes, as e.g., the GO biological processes cellchemotaxis (P-value = 6E−6), and the KEGG pathway endo-cytosis (P-value = 1E−6). The associated phenotypes [usingdata from the Mouse Genome Informatics (MGI) database]included inflammatory related phenotypes, e.g., decreased inflam-matory response (P-value = 3E−5), demyelination (P-value =3E−5), decreased macrophage cell number (P-value = 4E−5),and decreased tumor necrosis factor secretion (P-value = 8E−5).Furthermore, the tissue expression database shows a strong rolefor the nervous system, as the spinal nerve roots are the mostsignificant associated tissue (P-value = 9E−7). Finally, we usedGOEAST to investigate gene ontology (GO) terms, in whichthe genome-wide significant genes were overrepresented. Resultsfrom these analyses indicate significant enrichment for GO termsassociated with the glucose/insulin metabolism in the BiologicalProcesses category, e.g., negative regulation of insulin secretion (P-value = 2.39E−7, log odds = 4.46) and cellular response to glucosestimulus (P-value = 2.94E-11, log odds = 4.50). Other over-represented GO terms in the Biological Processes category wereglycolysis (P-value = 6.01E−7, log odds = 3.33), and skeletal mus-cle fiber development (P-value = 1.91E−5, log odds = 3.65). In theCellular Component category, we did not find highly significantGO terms. In the Molecular Function category the calcium-dependent protein serine/threonine phosphatase activity (P-value= 2.42E−12, log odds = 6.44), pyruvate kinase activity (P-value

= 1.62E−13, log odds = 6.32), and ferric iron binding (P-value =5.89E−8, log odds = 4.69) were most significant.

NETWORK ANALYSIS USING THE WISH NETWORK METHODGWA studies have several limitations, as discussed above: manySNPs are eliminated as they do not reach the strict genome-widesignificance thresholds and they do not take genetic networksbased on gene–gene interactions into account. We thereforeused the WISH network method to identify clusters of highlyinterconnected SNPs (modules), using the Duroc ∗ GöttingenMinipig (DM) intercross population. Detected modules were fur-ther investigated using various pathway detection approaches, toindicate their biological relevance.

Network constructionData reduction was based on the genome-wide significance(P-value < 0.05) and the connectivity of the SNPs, resultingin a selection of 2500 SNPs for network construction. Usingthe WISH network method based on genotype correlations, wedetected 17 modules of at least 50 SNPs per module. The moduleeigenSNP was calculated based on the first principal component,which explained 48–78% of the variation in the modules. We thenselected biologically interesting modules based on the Module-Trait Relationship (MTR), which was calculated as the correlationbetween the module eigenSNP and the traits of interest: the OIand 16 other obesity-related traits. Six modules were selectedfor downstream analysis based on the Genome-wide ModuleAssociation Matrix (GMAT), as they had a significant correla-tion with the OI, and a correlation of >0.4 with at least oneother obesity-related trait (Figure 5). Selected modules and theirMTR with OI were: Tan module (MTROI = 0.51), Lightgreenmodule (MTROI = 0.42), Lightyellow module (MTROI = 0.35),Purple module (MTROI = −0.33), the Royalblue module(MTROI = 0.21), and the Red module (MTROI = 0.19).

Downstream analysis of detected modulesGenes located within 20 kb of SNP present in selected moduleswere detected using the R-package NCBI2R. We performed path-way analysis using NCBI2R, GOEAST and GeneNetwork for allsix selected modules resulting from WISH network construction.

The Tan module included 84 SNPs, corresponding to 64 genes.GOEAST analyses resulted in the identification of several signifi-cant GO terms in the actin filament pathway (e.g., actin crosslinkformation, P-value = 2.59E−8) which is not directly associatedwith obesity. GeneNetwork analyses identified some overrepre-sented phenotypes related to diabetes (i.e., increased susceptibilityto autoimmune diabetes, P-value = 4.13E−4; and abnormal pan-creatic beta cell morphology, P-value = 6.58E−4), which may beresulting from the overrepresented Biological Process GO termbranched chain family amino acid metabolic process (P-value =6.98E−5), as those amino acids are associated with the metabolichomeostasis (Wang and Guo, 2013).

The Lightgreen module consists of 77 SNPs representing47 genes. The Molecular Function category in GOEAST showsone highly overrepresented GO term: purinergic receptor activity(P-value = 7.21E−25). Purinergic receptors have been implicatedin several different functions as, e.g., learning and memory, loco-motor and feeding behavior, and sleep. Two genes present in


http://www.genenetwork.nl




FIGURE 5 | Genome-wide Association Matrix (GMAT) of modules

detected using the WISH network method. On the y-axis the detectedmodules are visualized, and on the x-axis the different phenotypes (OI andbreeding values of 16 obesity-related phenotypes) are presented. In the

matrix the Module-Trait Relations (MTRs) are presented with the subsequentp-values. MTRs are colored red in case of a strong positive correlationbetween the Module and the Trait, and colored blue in case of a strongnegative correlation between the Module and the Trait.

this pathway are P2RX7 and ADORA2A. P2RX7 is encoding theprotein P2X purinoreceptor 7, which have been implicated in,e.g., neuronal cell death and inflammation (Skaper et al., 2010).Moreover, the P2X7 gene has been associated with diabetes, e.g.,by its influence on the regulation of beta cells (Glas et al., 2009).The ADORA2A (adenosine A2A receptor) gene is coding a pro-tein which is a receptor subtype for adenosine. It regulates bloodflow to the myocardium by vasodilation of the coronary arter-ies, potentially leading to hypotension. Besides, it has a neuronaleffect through its expression in the brain, associated with, e.g.,anxiety and depression (Ledent et al., 1997). A direct link withobesity or type 2 diabetes has not been shown, to our knowl-edge. Furthermore, the Biological Process category in GOEASTalso shows the overrepresented GO term fructose 2,6-bisphosphatemetabolic process (P-value = 3.95E−20), possibly because of thepresence of the PFKFB4 gene. PFKFB4 regulates the concen-tration of fructose 2,6-bisphosphate, which plays a key role inglycolysis (Uyeda et al., 1982).

The Lightyellow module consists of 72 SNPs representing 45genes. The GOEAST analysis only reveals some slightly signifi-cant overrepresented GO-terms, as, for example, the MolecularFunction alpha-glucosidase activity (P-value = 2.91E−7). TheGANC gene is present in this module and GO-term, encoding theneutral alpha-glucosidase C enzyme. This enzyme plays a majorrole in the glycogen metabolism and because of its effect on the

absorption of sugars from the gut, alpha-glucosidase inhibitorsare used in the treatment of type 2 diabetes (van de Laar, 2008).

The Purple module is the largest module, consisting of104 SNPs and representing 84 genes. The NCBI2R pathwayanalysis detected one, more general, pathway: metabolic path-ways (P-value = 0.06, not significant after multiple-testingcorrection). Using the GOEAST analysis, several GO-termswere detected as overrepresented, e.g., microtubule anchoring(P-value = 4.61E−31) and gamma-tubulin binding (P-value =1.81E−26).

The Royalblue module is the smallest module, consisting of57 SNPs representing 49 genes. The NCBI2R pathway analysiscame up with many significant overrepresented pathways, whichwere also significant after multiple-testing correction. Most sig-nificant were immune-related pathways, e.g., antigen processingand presentation (P-value = 7.6E−6) and herpes simplex infection(P-value = 1.85E−5), while others were more directly related tothe OOR diseases, e.g., type I diabetes mellitus (P-value = 5.8E−4).Using GOEAST several GO terms in the Biological Processes cat-egory were detected as overrepresented, e.g., collagen-activatedsignaling pathway (P-value = 2.03E−72) and smooth musclemigration (P-value = 4.06E−66).

Lastly, the Red module is consisting of 82 SNPs represent-ing 54 genes. GOEAST detected some overrepresented GO-terms associated with the monosialoganglioside sialytransferase






activity (P-value = 5.22E−12). The ST3GAL4 gene, present inthis module, has been associated with this GO-term. This genehas been associated to glycoprotein biosynthesis and cell surfaceglycobiology, which may be associated to the concentration ofliver enzymes which is often disturbed by diseases like obesity andgive an increased risk for, e.g., type 2 diabetes (Chambers et al.,2011).

Differentially wired (DW) networks of SNPsDifferentially wired (DW) or connected genes are genes whichare highly interconnected with other genes in one extreme sub-group (hub genes), while lowly interconnected with other genesin the other extreme subgroup. Here, we examined the differen-tial connectivity (k_diff) of SNPs between the lean subgroup andthe obese subgroup. These DW genes in one group vs. the otherwould be biologically interesting candidates as they might pointto differences in underlying genetic regulation for manifestationof the obesity or leanness. In total, 55 SNPs showed an absolutedifferential connectivity above 0.6, which resulted in the detec-tion of 36 genes (including 10 uncharacterized proteins) withina flanking distance of 20 Kb. We here present the genes stronglyassociated with obesity or obesity-related diseases before, all DWSNPs and corresponding genes are presented in Additional File 2.

Several DW genes were hub genes in the lean sub network,but had a low interconnectivity in the obese sub network, e.g.,UBR1, PNPLA8, and CTNAP2. The UBR1 gene (k_diff = 0.62)encodes a member of the E3 ubiquitin ligase family, which func-tions in the N-end rule pathway. This pathway has been associatedwith various functions, e.g., the control of apoptosis (Ditzel et al.,2003). Moreover, Ubr1 knockout mice also demonstrate sub-tle effects in muscle protein degradation and fat metabolism(Kwon et al., 2001). The PNPLA8 gene (k_diff = 0.62) encodesa member of the patatin-like phospholipase domain containingproteins. These phospholipases catalyze the cleavage of fatty acidsfrom membrane phospholipids. The CNTNAP2 gene (k_diff =0.62) encodes contactin-associated protein-like 2, a protein whichfunctions in the nervous system as cell adhesion molecules andreceptor. A neuronal impact on obesity is suggested by this gene,by the influence on the potassium channel at the nodes of Ranvier,resulting in the regulation of diet-induced obesity (Buchner et al.,2012).

On the other hand, DW genes that were detected with alow interconnectivity in the lean sub network were in fact hubgenes in the obese sub network. Two of these DW genes havepreviously been reported to be associated with obesity and dia-betes, respectively: the NUCB2 gene and the BCL11A gene. Thenucleobindin-2 (NUCB2) gene (k_diff = −0.80) is a precursor ofnestafin-1, a hypothalamic anorectic neuropeptide. It has recentlybeen discovered that this neuropeptide is expressed in, e.g., pan-creatic islet cells and the central nervous system. It seems toplay an important role in hypothalamic pathways regulating foodintake and energy homeostasis, and it has been shown to play animportant role in regulation of food intake in obese individuals(Abaci et al., 2013). Moreover, it has been shown that nesfatin-1is expressed in the same human gastric X/A-like cells as ghre-lin, a hunger-stimulating hormone. The expression of nestafin-1and ghrelin is differentially regulated under obese conditions:

nestafin-1 increases and ghrelin decreases with an increasing BMI,toward a further adaptive change that may counteract furtherbody weight increase (Stengel et al., 2013). The BCL11A gene(k_diff = −1.00) encodes the B-cell lymphoma/leukemia 11Aprotein. The corresponding mouse gene has been associated withleukemia, through its interaction with BCL6. However, this genehas also been detected in GWA studies to be associated with type 2diabetes and pancreatic β-cell function (Simonis-Bik et al., 2010;Langberg et al., 2012).

DISCUSSIONIn this systems genetics analysis using an F2 porcine modelof OOR diseases, we used several methods to identify under-lying genetic variants, modular networks, biologically relevantpathways, and hub genes. All these analyses were based on theOI, an aggregate genetic value constructed by combining theestimated breeding values of nine different obesity-related pheno-types, thought to be highly relevant in defining both overall andvisceral obesity. We used the principles of genetic selection indexmethod often used in animal breeding to construct an aggregategenetic index for obesity in porcine model. To our knowledge,this is the first study to develop such an aggregate (genetic) OI forsystems- and network genetics investigations. The porcine modelwas subject to extensive phenotyping, which is either not possi-ble or expensive in human populations. Previous investigationshave shown that a large proportion of variation in these pheno-types is genetically determined (Kogelman et al., 2013), therebyoffering an opportunity to exploit this resource to provide novelinsights into the genetic complexity of the disease. By combininga key number of the obesity phenotypes into one, we were ableto investigate this complex disease with one simplified aggregatephenotype. Moreover, the OI allows selection of extremely obeseand extremely lean animals, resulting in a greater power to detectgenes influencing obesity.

GWA STUDY AND VALIDATION OF IDENTIFIED REGIONSA single SNP association approach was implemented via GWAanalysis on the OI, and subsequently seven selected GWAS regionswere validated using combined LDLA, and further investigatedby analyzing haplotype blocks in these regions using Haploview.A large number of SNPs (representing 289 genes) were foundto be significantly associated with OI via GWA. The high num-ber of genome-wide significant SNPs is likely due to the use ofthe OI that combines nine different phenotypes. Consequently,we do expect to find a greater number of SNPs for this aggre-gate phenotype than for a situation where only one phenotypeis analyzed. Moreover, since OI represents only the polygenetic(or estimated breeding) value for each phenotype used in its con-struction, it is likely that re-regressing on a regressed “phenotype”will yield inflated estimates of the regression coefficient. Thisexplains the high number of significant SNPs in the GWA study,but also the very small p-values obtained via the LDLA approach.However, since regression was used only to rank SNPs for datareduction, and regression coefficients were not used for networkconstruction, this did not influence the network analysis.

The resource population was constructed by crossing genet-ically divergent breeds differing with respect to obesity related




EDU

Highlight

EDU

Highlight

EDU

Highlight


phenotypes, followed by sibling mating. The resultant F2 porcineintercross population allows Linkage Disequilibrium (LD) toexploit family information to validate GWA findings. TraditionalGWA studies rely solely on across population LD that onlyextends to short genetic intervals. Consequently, causal variantsare hard to identify if they are not located in the immediate vicin-ity of the markers assayed in a study. Since within family LDextends to larger genetic intervals, analytical approaches that usethis information along with across family LD (e.g., combinedlinkage disequilibrium linkage analyses; LDLA) provides greaterpower and finer resolution of the candidate loci particularly ingenomic regions with low population wide LD (Meuwissen et al.,2002). LDLA is a fine mapping approach that combines bothwithin-family linkage and population-wide LD in one analysis.Since single marker GWA analysis is based on population wideLD and linkage information was available given the 3-generationF2 intercross pedigree, it was a reasonable further step to capi-talize on this F2 pedigree structure to validate findings of highlysignificant regions in the GWA analysis.

Obesity is a very complex trait, with many different involvedbiological pathways, and consequently a huge number of geneswill be related. Many SNPs showed a significant association withOI, and using a flanking distance of 20 Kb, we consequently foundmany genes located in or nearby those SNPs. As discussed, wewould expect that SNPs over a rather long distance will be inLD, as a consequence of using parental lines genetically diver-gent for two different characteristics (leanness vs. obese) each linehaving larger LD blocks passed onto F2 pigs. This would meanusing a larger flanking distance, consequently detecting even moregenes in the regions of associated SNPs. However, we have seen inthis study that there are rather small haplotype block sizes, andbecause of the high number of SNPs detected, we limited genedetection to a flanking distance of 20 Kb.

As expected, several of our GWA study finding could bedirectly or indirectly linked to obesity. The number of associatedgenes is also in agreement with other studies: GWA studies alonehave indicated at least 37 genes related with BMI, 14 genes withWHR adjusted for BMI, 3 genes with fat percentage, etc. (Fall andIngelsson, 2014). Only one of our GWA study findings overlapwith previous findings, published in Fall and Ingelsson (2014), asfor example the MAP2K5 gene (P-value = 2.32E−9) on chromo-some 1 of the pig genome. This gene has previously been detectedby GWA studies on BMI by Speliotes (Speliotes et al., 2010) andWen (Wen et al., 2012). However, many loci previously reportedto be associated with obesity were not detected in our study, e.g.,FTO and MC4R gene. SNPs located in or near both these genesdid not survive data quality control. Moreover, many SNPs couldnot be annotated, mainly due to the limited annotation of the piggenome, which is also a main limitation in the pathway detec-tion among SNPs detected using the GWA study. Pathway analysisdid show some biologically relevant pathways, e.g., negative regu-lation of insulin secretion, but unfortunately, the findings of thedifferent pathway methods did not overlap completely. The inad-equate knowledge about the annotation of the pig genome alsoresults in limited knowledge about biological pathways presenton the pig genome. We therefore chose to identify the genes inclose proximity of SNPs, and use those genes in a human pathway

setting, by using the human reference genome and annotationinformation. Furthermore, the variation in results between thedifferent pathway methods is also due to the different referencedatasets, as for example GOEAST works with GO terms only, andthe NCBI2R method was based on KEGG pathways.

NETWORK APPROACHWe used the WISH network method to cluster SNPs based on thecorrelation patterns of genotypes. This systems genetics approachgave us the opportunity to analyze the interactions betweenthousands of SNPs, and thereby overcome major limitationsof GWA studies. GWA studies use a very stringent genome-wide significance threshold, excluding potentially biologicallyrelevant associated SNPs with very small effect sizes. Here, weselected SNPs based on a less stringent nominal P-value threshold(0.05), increasing the probability of including SNPs with smallereffect sizes that are biologically relevant for the trait of interest.Moreover, we are not only interested in the detection of singlegenetic variants associated with obesity, but also in the detectionof molecular pathways through which the genetic variants exertstheir effects. Because of the exclusion of SNPs with interactioneffects on phenotypes in GWA study, the elucidation of importantpathways is limited. By using a WISH network method, we areable to include and analyze genome-wide interactions betweenSNPs and relate them to molecular and cellular functions.

WISH network construction was based on genotypic correla-tion and since genotypic correlations largely represent LD, partof the SNPs in the network modules were found to be in closephysical proximity. However, many SNPs in the modules were notphysically co-located. For example, the Tan module has the high-est correlation with OI and consists of 84 SNPs. Of these, 9 SNPscould not be mapped to the porcine genome and the remain-ing 75 SNPs were distributed over all porcine chromosomes,except chromosome 15. As shown in Table 2, there is a fairlyequal distribution of SNPs included in this module over differ-ent porcine chromosomes. Moreover, in cases where many SNPsare co-located on one chromosome, they cover a large area of thechromosome. This indicates that the WISH network constructionbased on genotypic correlations does not only capture SNPs phys-ically co-located. However, previous studies have also shown thatvery distant loci located on different chromosomes can be in LDwith each other (Flint-Garcia et al., 2003). Studies have advocatedinvestigating LD between loci, as a SNP highly associated withthe disease could be in LD with a causal SNP (Weiss and Clark,2002). By detecting clusters of SNPs that are in LD with eachother, and determining their functional annotation, we attemptto investigate the biological relevance of SNPs in genome-wideLD. The WISH network method may also be applied using anepistatic interaction model, using the regression coefficients of theSNPi

∗SNPj interaction term. However, as discussed previously,regression models using a regressed phenotype like OI will likelyyield inflated coefficients. Therefore, we did not apply the WISHnetwork method based on epistatic interactions in this study.

In total six WISH network modules were further examinedusing different pathway analysis methods. Those methods showdifferent results which could be directly or indirectly related toobesity and obesity related diseases. As expected, many pathways






Table 2 | Distribution of SNPs over the chromosomes in the Tan

Module using the WISH network method based on genotypic

correlations.

Chromosome # SNPs Location

Minimum Maximum

1 17 87.621.858 291.903.522

2 11 2.338.853 145.102.337

3 3 74.111.456 141.121.427

4 7 21.818.546 142.317.862

5 3 4.874.980 10.686.539

6 4 43.090.860 126.991.726

7 4 87.893.776 97.500.922

8 2 20.144.538 109.213.653

9 4 48.813.606 132.656.946

10 1 58.238.109 58.238.109

11 4 37.160.369 82.659.616

12 2 52.678.657 54.767.975

13 5 31.119.717 172.601.927

14 2 62.514.781 63.414.448

16 3 8.438.953 68.583.123

17 1 9.342.342 9.342.342

X 1 6.909.025 6.909.025

were related to metabolic processes (e.g., fructose 2,6-bisphosphatemetabolic process and branched chain family amino acid metabolicprocess). The variety of results show the complexity of the diseaseunder study, as many different metabolic associated pathways andGO terms show up in the different modules. To further investigatethe relation with the insulin pathway, and shown the additionalvalue of systems genetics approaches, we looked further into theconnectivity’s of the SNPs present in the Lightgreen Module, asthis one showed associations with, e.g., the glycolysis pathway.Three out of the seven highest connected SNPs were located inthe FRMD4B, ADAMTS9, and XCR1 genes. Those genes were allpreviously related to the insulin pathway or diabetes, where XCR1(encoding a chemokine receptor) even linked obesity to insulinresistance before (Ota, 2013). The FRMD4B gene is part of theGRP1 signaling complex, thereby recruited in response to insulinreceptor signaling, but genetic variants have also been associ-ated with heart failure (Matkovich et al., 2010). The ADAMTS9gene has been associated to Type 2 Diabetes in a meta-analysis ofgenome-wide association data (Zeggini et al., 2008). Those genesshow in the network construction a perfect correlation (of 1) witheach other, and as they lie on the porcine genome in a region ofapproximately 18 Mb of each other, it is likely that they are withinthe same haplotype block. However, the integrative systems genet-ics approaches also shows the potential of further investigatingthose genes, and gaining knowledge about the genetic architectureof complex traits. We examined the differential wiring of SNPsbetween two subgroups that are genetically lean and obese. TheseDW genes in one group vs. the other would be biologically inter-esting candidates as they might point to differences in underlyinggenetic regulation for manifestation of the obesity or leanness.We detected 55 DW SNPs and 36 co-located genes within 20 kb

(e.g., UBR1, PNPLA8, and CTNAP2), which have significantimplications for development of obesity and or obesity-relateddiseases, as shown in additional file 2. The WISH network wasable to identify GO terms and pathways which were not identi-fied by pathway analysis of the GWA study results, and moreover,none of the DW genes were identified using the GWA study.

In general, the study describes several obesity related genesand pathways that accord with the complexity of the disease. Asthis F2 pig resource population is extensively phenotyped, fur-ther studies could yield novel biological insights underlying theassociation between phenotypes used in this study and the identi-fied genes. The results of this study overlap with previous humanfindings, while also identifying novel genes not previously knownto be associated with human obesity. However, further valida-tion of those loci will be needed to confirm the association withobesity, and their exact function in biological pathways. Once val-idated, these findings could potentially be extended to humans inorder to improve the treatment of obesity and eventually reducecomplementary problems resulting from obesity.

CONCLUSIONThis is the first study to develop an aggregate (genetic) OI basedon the principles of quantitative genetics and animal breeding,to study genetic of obesity. We demonstrated the potential ofnetwork-based systems genetics approaches to reveal biologicaland genetic background of complex phenotypes that is other-wise not identified via traditional genetics/genomic analyses. Herewe performed GWA analysis, validated these results via LDLAand evaluation of the haplotype block sizes, and subsequentlyperformed enrichment analysis to identify biological pathwaysassociated with obesity and obesity related traits. Subsequently,scale-free WISH networks were constructed and different clus-ters of highly interconnected SNPs that were putatively relatedto obesity and related diseases were identified, demonstrating theimportance of genetic interactions in obesity. We have also exam-ined the differential connectivity or wiring of SNPs between twosubgroups that are genetically obese and lean and detected 36co-located genes, which have significant implications for develop-ment of obesity and or obesity-related diseases. Many genes withdiverse functions and consequently many different pathways wereidentified by WISH and differential wiring approaches that werenot detected by traditional GWA analyses, thereby demonstratingthat integrative systems approaches could potentially yield novelinsights into the genetic determination of obesity and its relationto other diseases. This study, to the best of our knowledge, is thefirst network-based systems genetics analyses on an experimentalpig population in which a wide range of obesity traits have beeninvestigated, and reaffirms the complexity associated with obesityas a disease.

MATERIALS AND METHODSThe complete workflow, from materials to results, is presentedin Figure 6. The core materials for this study are the purpose-built F2 pig model for obesity, high-throughput genotypic data,and nine selected obesity phenotypes. Using these materials, wedeveloped first an aggregate (genetic) OI and applied GWASto the OI. The outputs of GWAS were then used in WISH




EDU

Highlight

EDU

Highlight

EDU

Highlight


FIGURE 6 | Workflow visualizing the different methods to detect obesity-related genes and pathways.

network construction. The LDLA methods were used validatetop genomic regions before conducting pathway, network andfunctional annotation analyses. These approaches then resultedin identification of novel obesity-related genes and pathwayswhich provide deeper understanding of genetic control of obesitydevelopment.

THE F2 PIG RESOURCE POPULATION: PHENOTYPE AND GENOTYPINGAn F2 pig resource population, genetically divergent for OORtraits, was established as described previously (Kogelman et al.,2013). Briefly, the F2 pig resource population was created usingDanish production pig breeds, i.e., purebred Yorkshire (YY) andDuroc (DD) sows from a DanBred breeding herd and GöttingenMinipig (MM) boars from Ellegaard A/S in the parental gen-eration. The production pigs have been selected for leannessand growth among other traits during the last 60 years, whereasGöttingen Minipigs are bred principally for their small size andease of handling and are prone to obesity. They are also known toshare the metabolic impairments seen in obese humans (Wangand Guo, 2013). The F2 pig resource population consists of24 parental animals, 78 F1- and 454 F2-animals. The resultingpopulation of 556 animals can be divided into Duroc∗Göttingen

Minipig (DM, 279 pigs) and Yorkshire∗Göttingen Minipig (YM,277 pigs) subpopulations. Animal care and maintenance wasconducted according to the Danish “Animal Maintenance Act”(Act 432 dated 09/06/2004). This F2 pig resource was exten-sively phenotyped for several OOR traits that included weight,conformation, dual energy x-ray absorptiometry (DXA) scan-ning and slaughter measurements. Genetic parameters estimatedvia this F2 pig resource population revealed normally distributedgenetic values within those OOR traits, and the potential for fur-ther genomic and system genetic investigations (Kogelman et al.,2013). Descriptive statistics and heritabilities of the main traits ofinterest in this study are presented in Table 3.

Blood was collected from all pigs in the F2 pig resourcepopulation from the jugular vein. Genomic DNA was extractedfrom EDTA stabilized blood using a simple salting out procedure(Miller et al., 1988). Genotyping was performed by GenoSkanA/S, Tjele, Denmark using the Illumina 60K porcine SNP-chip.Quality control (QC) was performed in the R-package GenABEL(Aulchenko et al., 2007), resulting in the exclusion of seven ani-mals due to low call rate (≤0.05), and seven animals due to toohigh identical by state (IBS ≥ 0.95). Furthermore, 3240 markerswere excluded because of a low call rate (≥0.05), 7615 markers






Table 3 | Descriptive statistics and heritabilities of main

fatness-related traits in the pig resource population, as published in

Kogelman et al. (2013).

Trait Unit Abbreviation N Mean SD h2

WEIGHT

2 months Kg WT2m 439 12.45 4.52 0.78

7 months Kg WT7m 405 94.45 17.51 0.39

Average daily gain Kg/day ADG 403 0.44 0.08 0.54

Abdominal

circumference

Cm ABD7m 404 122.7 10.56 0.28

BMIa Kg/cm2 BMI7m 403 132.86 20.71 0.23

DXA

Fat Kg DXAfat 438 22.92 8.53 0.43

Lean Kg DXAlean 438 100.42 37.33 0.71

Mass Kg DXAtotal 438 123.34 44.93 0.67

% fat % DXA%fat 438 18.64 2.87 0.57

Fasting glucose Mmol/L FGL 146 4.50 2.28 0.49

SLAUGHTER

Carcass weight Kg SLcw 358 56.33 11.77 0.54

Meat percentage % SL%meat 330 43.39 6.65 0.18

Weight leaf fat Kg SLfat 396 2.59 1.10 0.23

Backfat 1b Mm SLbf1 330 31.90 8.54 0.22

Backfat 2c Mm SLbf2 330 35.44 9.12 0.23

Omental fat Kg SLfat_om 257 352.72 145.28 0.52

Intestinal fat Kg SLfat_int 219 19.88 8.32 0.08

aBMI, body mass index, calculated as [weight/(length)2].bBackfat measured between third and fourth lumbar vertebra, 8 cm off midline.cBackfat measured between third and fourth last rib, 8 cm off midline.

were excluded because of their low minor allele frequency (MAF≤ 0.05), and 4723 markers were excluded as they were not inHardy–Weinberg equilibrium (P-value < 1E−5). This resulted inthe analysis of 40,910 markers and 538 animals.

OBESITY INDEX BASED ON BREEDING VALUESTo maximize genetic progress, selection indices are extensivelyused in animal breeding (Cameron, 1997) to select animals withdesirable genotypes for a particular phenotype of interest. Theselection index theory was used to create one genetic value repre-senting the degree of obesity for all animals, resulting in one moredistinct phenotype for obesity. Using the estimated variance com-ponents, combining several OOR traits into one aggregate geneticvalue for all animals in the F2 pig resource population, we cre-ated the OI. Traits used for construction of the OI were: weightat slaughter age (WT7m), abdominal circumference at slaugh-ter age (ABD7m), average daily gain (ADG), estimated fat massat DXA (DXAfat), estimated percentage of fat at DXA scanning(DXA%fat), backfat thickness at position 1 (SLbf1) and position 2(SLbf2), weight of leaf fat at slaughtering (SLfat), and omental fatat slaughtering (SLfat_om). These phenotypes were selected to col-lectively represent and be associated with “obesity” as an excessiveamount of adipose tissue.

By combining the estimated breeding or genetic values (EBVs)for several traits per animal, and weighting individual geneticvalues of each traits by their relative biological weight (v), one

aggregate genetic value per animal is calculated representing allphenotypes as:

I = b1x1 + b2x2 + · · · + bjxj = b′x

Where I is the selection index, bj is the selection index coeffi-cient (weight) for the jth observation, and xj is the jth phenotypicobservation. The selection index coefficient b is calculated basedon the phenotypic and genetic (co-)variance components, multi-plied by a biological assigned weight for the particular traits:

b = P−1Gv

Where P−1 is the inverse matrix of the phenotypic (co-)variances,G is the matrix of the genotypic (co-)variances and v is a vectorwith the biological assigned weights. The P and G matrices areconstructed as follows:

P =

⎡⎢⎢⎣

σ 2P1 σP1,2 . . . σP1,9

σP2,1 σ 2P2 . . . σP2,9

. . . . . . . . . . . .

σP9,1 σP9,2 . . . σ 2P9

⎤⎥⎥⎦ , G =

⎡⎢⎢⎣

σ 2G1 σG1,2 . . . σG1,9

σG2,1 σ 2G2 . . . σG2,9

. . . . . . . . . . . .

σG9,1 σG9,2 . . . σ 2G9

⎤⎥⎥⎦

where σ 2Pi is the phenotypic variance of the ith trait, σPi,j is

the phenotypic covariance between the ith and jth trait, σ 2Gi is

the genotypic variance of the ith trait, and σGi,j is the geno-typic covariance between the ith and jth trait. The phenotypicand genetic (co-)variance components were estimated using aseries of bivariate animal models for all combinations betweenselected traits for the OI, as presented in our previous study(Kogelman et al., 2013); all models were implemented usingASReml (Gilmour et al., 2009).

As described earlier, each trait was assigned a biological weight(v) based on biological assumptions that was used to calculate theselection coefficient: weight at 7 months of age (v = 0.1), abdom-inal circumference at 7 months of age (v = 0.1), average daily gain(v = 0.1), body fat estimated by DXA scanning (v = 0.5), per-centage of body fat estimated by DXA scanning (v = 0.5), weightof leaf fat at slaughter (v = 0.8), back fat thickness at position 1(v = 1), back fat thickness at position 2 (v = 1), and weight ofomental fat at slaughter (v = 0.8).

Accordingly, OI is defined as:

OI = bWT7m ∗ XWT7m + bABD7m ∗ XABD7m + bADG ∗ XADG

+ bDXAfat ∗ XDXAfat + bDXA%fat ∗ XDXA%fat + bBF1 ∗ XBF1

+ bBF2 ∗ XBF2 + bSLfat ∗ XSLfat + bSLfat_om ∗ XSLfat_om

where OI is the Obesity Index, b is the selection index coefficientcalculated using the (co-)variance components and biologicalweight (v), and X is the estimated breeding value of the selectedtraits. The estimated breeding values, representing the animalsdeviation from the mean of the population, were estimatedusing the variance component estimation models presented inKogelman et al. (2013) using ASReml (Gilmour et al., 2009).





GENOME-WIDE ASSOCIATION ANALYSISThe R package GenABEL (Aulchenko et al., 2007) was used totest the association between individual SNPs distributed through-out the porcine genome, and the OI using the complete F2pig resource population. Since the OI is derived from estimatedgenetic (breeding) values calculated after correcting for popula-tion structure and all other environmental and fixed effects (seeKogelman et al., 2013) a simple regression of the OI on SNPgenotypes was used to perform GWA analysis to avoid double-or over-correction. The basic linear model was:

y = μ + g + e

where y = OI, μ = the phenotypic mean, g = the SNP genotypes(coded as 1 and 2, 0 for missing), and e = the model errors.We calculated the Bonferroni corrected p-values by dividing theresulting p-value by the total number of SNPs passing QC thresh-olds (as described above), resulting in a suggestive association atPadj = 1.22E−6 (0.05/number of SNPs) and a highly significantp-value at Padj = 2.44E−8 (0.001/number of SNPs).

VALIDATION OF GWA STUDY FINDINGS BY LDLA AND HAPLOTYPEBLOCK ANALYSESThe genomic regions with most highly genome-wide significanceidentified by GWA study (Padj = 2.44E−8) were further vali-dated using the combined LDLA approach (Meuwissen et al.,2002). Regions were selected when they were in the top 10 GWAresults (most highly significant) and the associated gene had abiological role (in-) directly associated to obesity. Length of theregions was arbitrary taken, in such a way that several SNPsup- and downstream were investigated. Statistical significanceusing the LDLA approach was calculated via a likelihood ratiotest of the full model (OI regressed over the phenotypic meanand Identity by Descent probabilities of chromosomal segmentsflanked by successive marker pairs) vs. the null model (containingonly the phenotypic mean). The Identity by descent probabilitieswere estimated using a linkage disequilibrium multilocus iter-ative peeling (LDMIP) algorithm described in Meuwissen andGoddard (2010). Furthermore, we identified haplotype blocks inthe genomic regions that were detected as being highly significantby both GWA and LDLA analyses, as they are likely to be the mostpromising genomic regions harboring genetic variants affectingOI. The structure of haplotype blocks in this F2 pig resource pop-ulation was determined and plotted using Haploview software(v4.2) (Barrett et al., 2005). Pairs were defined as being in “strongLD” using an upper confidence bound on D′ > 0.98 and a lowerconfidence bound on D′ < 0.7.

WISH NETWORK CONSTRUCTION AND ANALYSISWe previously published the WISH network method based usingwhole genome genotype data, giving the opportunity to iden-tify clusters of highly interconnected SNPs (modules) and relatethem to phenotypes (Kogelman and Kadarmideen, 2014). Weapplied these methods to construct WISH networks based ongenotypic correlations, using the pipeline as presented previously,in order to identify biologically relevant pathways underlyingobesity and obesity related traits in a subpopulation of the pig

resource population. The QC of the high-throughput genotypedata was performed with the same parameters as for GWA study,resulting in 40,194 SNPs and 266 pigs from the Duroc∗GöttingenMinipig intercross.

For network construction all SNPs with a genome-wide sig-nificance below 0.05 were selected for network construction,resulting in the selection of 9485 SNPs. Animals were selectedbased on their OI: 75 animals were selected based on an extremeOI (25 low OI, 25 intermediate OI, and 25 high OI). These selec-tions resulted in a 75 ∗ 9485 matrix of the genotypes coded as1,2,3 for each animal. Because of computational limitations, thesize of the data set was further reduced by selecting SNPs based ontheir connectivity, which is the sum of the connection strengthsof a particular SNP with all other SNPs. Genes with a high con-nectivity, also called hubgenes, are thought to be biologicallyimportant and therefore, only the top 2500 SNPs were selected(normalized connectivity > 0.12). To pursue scale-free topology,a power γ was chosen in such a way that the R2 (the scale-freetopology index) approaches one. A power γ of 5 resulted in anR2 of 0.88, which was used to create an adjacency matrix by cal-culating the Pearson’s correlations among SNPs and raising thisto the power γ of 5. The network was constructed based onthe topological overlap measure (TOM) between SNPs, where ahigh TOM represents a high share of neighbors between a pairof SNPs, and consequently, a low TOM represents a low share ofneighbors between a pair of SNPs. Based on the TOM, clusters ofhighly interconnected SNPs were detected, called modules, usingthe Dynamic Tree Cutting algorithm (Langfelder et al., 2008).Modules were selected for downstream analysis based on theirModule-Trait Relation (MTR), which is calculated by correlatingthe module eigenSNP (the first principal component, explain-ing most of the variance in the module) with the OI and otherOOR traits. P-values were represented by the Student asymptoticp-value for the given correlations. Other OOR traits were rep-resented by the estimated breeding values, previously calculated(Kogelman et al., 2013). Modules with a significant correlation(P-value < 0.001) with the OI and an MTR > 0.4 with at oneother OOR trait were selected for pathway analysis.

Differential wiring of SNPs in obesity (DW networks)It is expected that biologically important pathways related tothe trait under investigation, will show a different activity pat-tern between the two extreme groups. In other words, a path-way could be induced in the case-group, while repressed in thecontrol-group. Highly interconnected SNPs, and consequentlytheir mapped genes, are called hub genes, which are poten-tially biologically relevant genes. Therefore, we investigated whichSNPs were DW (connected) between the lean and obese animals.The interconnectivity is represented by the sum of correlations ofa particular SNP with all other SNPs within a (obese or lean) sub-network. First, the 50 most extreme lean animals and the 50 mostextreme obese animals were selected based on the OI. Based onthe same selected SNPs (n = 40, 194) as in the normal networkconstruction the connectivity of all SNPs in the lean (k_lean) andobese (k_obese) subnetwork was calculated. Then, the differen-tial wiring or connectivity (k_diff) was calculated by subtractingk_obese from k_lean. This resulted in a positive k_diff for SNPs






that had high interconnectivity in the lean subnetwork, but lowinterconnectivity in the obese subnetwork. Subsequently, negativevalues for k_diff were found for SNPs that had high interconnec-tivity in the obese subnetwork, but low interconnectivity in thelean subnetwork. SNPs with an absolute k_diff above 0.6 wereselected for further investigation.

PATHWAY ANALYSISSNP selection and gene detectionSNPs that were cross-validated as being highly significant byboth GWA and LDLA study and those SNPs that were presentin the modules as being highly associated with the OI fromWISH network construction were all selected for further path-way profiling analyses. Genes mapping to the detected SNPs wereobtained using Biomart (Ensembl v73) (Haider et al., 2009).All identified genes were used for pathway analysis. Secondly, tocover the promoter region of genes, which lay outside positionscovered by Biomart, we used the NCBI2R R-package (Melvilleand Melville, 2012) (available at http://cran.r-project.org/web/packages/NCBI2R/index.html) which uses a list of SNPs as aninput and gives (if present) the genes and their annotation inthe indicated region. We used a flanking distance of 20 kB, as theaverage gene size is 30 kB and we used an extra 5 kB to cover thepromoter region.

Pathway detectionGenes located within 20 kb of SNPs identified via GWA andWISH network analyses (modules) were used for gene enrich-ment analysis of biological pathways in order to identify thepotential biological relevance of detected SNPs. Since path-way analysis is very dependent on the databases used forbiological annotation, we used different tools that leverage infor-mation available in different publically available tools to comple-ment each other: NCBI2R (available at http://cran.r-project.org/web/packages/NCBI2R/index.html), Gene Ontology EnrichmentSoftware Toolkit (GOEAST) (Zheng and Wang, 2008) andGeneNetwork (http://www.genenetwork.nl).

Using the R-package NCBI2R, pathways in the identified geneswere detected using the GetPathways() function. The same func-tion was used on all genes present in and around (flankingdistance = 20 Kb) the SNPs which passed QC, as a reference path-way set. The significance level of present pathways was calculatedusing the Fisher’s exact test and multiple-testing correction wasapplied using the Bonferroni-correction. Secondly, GOEAST wasused to identify overrepresented GO terms among the identifiedgenes. The Gene Batch tool in GOEAST was used to import thegene symbols and to identify significantly overrepresented GOterms and corresponding pathways were visualized (Zheng andWang, 2008). Thirdly, GeneNetwork (http://www.genenetwork.

nl) was used to identify overrepresented GO terms, KEGG path-ways, phenotypes, and tissues. GeneNetwork is constructed usinghuman, mouse, and rat expression data, to predict gene func-tions against known pathways and gene sets in various biologicaldatabases. Overrepresentation of GO-terms and pathways wastested within the GeneNetwork tool, using the Mann–Whitney Utest, and P-values were afterwards corrected for multiple testingusing the Bonferroni correction.

AUTHOR CONTRIBUTIONSHaja N. Kadarmideen was the project leader and contributed todesigning quantitative- and systems genetics analyses includingsingle SNP association methods, obesity index, network construc-tion and pathway profiling approaches, and supervised Lisette J.A. Kogelman in these analyses. Merete Fredholm contributed todesigning the F2 resource population and supervised collectionof biological material and phenotypic measurements on all pigs.Lisette J. A. Kogelman analyzed all the data. The LDLA analysiswas performed by Sameer D. Pant. Lisette J. A. Kogelman wrotethe first draft of the manuscript. All authors wrote, read, andapproved the final version of the manuscript.

ACKNOWLEDGMENTSThe project is supported by a grant from the Ministry ofScience and Technology to the “UNIK Project for Food Fitnessand Pharma for Health,” funding from the Danish Council forStrategic Research to BioChild Project (www.biochild.ku.dk),and from a Ph.D. stipend awarded to Lisette J. A. Kogelmanfrom University of Copenhagen. Authors thank EU-FP7 MarieCurie Actions—Career Integration Grant (CIG-293511) grantedto Haja N. Kadarmideen for funding this study. Authors thankTheo Meuwissen for providing LDLA software for our analyses.

SUPPLEMENTARY MATERIALThe Supplementary Material for this article can be foundonline at: http://www.frontiersin.org/journal/10.3389/fgene.2014.00214/abstract

REFERENCESAbaci, A., Catli, G., Anik, A., Kume, T., and Bober, E. (2013). The relation of serum

nesfatin-1 level with metabolic and clinical parameters in obese and healthychildren. Pediatr. Diabetes 14, 189–195. doi: 10.1111/pedi.12009

Aulchenko, Y. S., Ripke, S., Isaacs, A., and Van Duijn, C. M. (2007). GenABEL: anR library for genome-wide association analysis. Bioinformatics 23, 1294–1296.doi: 10.1093/bioinformatics/btm108

Barrett, J. C., Fry, B., Maller, J., and Daly, M. J. (2005). Haploview: analysisand visualization of LD and haplotype maps. Bioinformatics 21, 263–265. doi:10.1093/bioinformatics/bth457

Bener, A., Zirie, M., and Al-Rikabi, A. (2005). Genetics, obesity, and environmentalrisk factors associated with type 2 diabetes. Croat. Med. J. 46, 302–307.

Brazhnik, P., De La Fuente, A., and Mendes, P. (2002). Gene networks: how to putthe function in genomics. Trends Biotechnol. 20, 467–472. doi: 10.1016/S0167-7799(02)02053-X

Buchner, D., Geisinger, J., Glazebrook, P., Morgan, M., Spiezio, S., Kaiyala, K.,et al. (2012). The juxtaparanodal proteins CNTNAP2 and TAG1 regulatediet-induced obesity. Mamm. Genome 23, 431–442. doi: 10.1007/s00335-012-9400-8

Cameron, N. D. (1997). Selection Indices and Prediction of Genetic Merit in AnimalBreeding. Edinburgh: Cabi.

Chambers, J. C., Zhang, W., Sehmi, J., Li, X., Wass, M. N., Van Der Harst,P., et al. (2011). Genome-wide association study identifies loci influencingconcentrations of liver enzymes in plasma. Nat. Genet. 43, 1131–1138. doi:10.1038/ng.970

Civelek, M., and Lusis, A. J. (2014). Systems genetics approaches to understandcomplex traits. Nat. Rev. Genet. 15, 34–48. doi: 10.1038/nrg3575

Comuzzie, A. G., Cole, S. A., Laston, S. L., Voruganti, V. S., Haack, K., Gibbs,R. A., et al. (2012). Novel genetic loci identified for the pathophysiologyof childhood obesity in the hispanic population. PLoS ONE 7:e51954. doi:10.1371/journal.pone.0051954

Cordell, H. J. (2009). Detecting gene-gene interactions that underlie human dis-eases. Nat. Rev. Genet. 10, 392–404. doi: 10.1038/nrg2579


http://cran.r-project.org/web/packages/NCBI2R/index.html







www.biochild.ku.dk






Cunha, P., Back, L., Sereia, A., Kubelka, C., Ribeiro, M., Fernandes, B., et al.(2013). Interaction between obesity-related genes, FTO and MC4R, associ-ated to an increase of breast cancer risk. Mol. Biol. Rep. 40, 6657–6664. doi:10.1007/s11033-013-2780-3

Diez, D., Wheelock, A. M., Goto, S., Haeggstrom, J. Z., Paulsson-Berne, G.,Hansson, G. K., et al. (2010). The use of network analyses for elucidatingmechanisms in cardiovascular disease. Mol. Biosyst. 6, 289–304. doi: 10.1039/b912078e

Ditzel, M., Wilson, R., Tenev, T., Zachariou, A., Paul, A., Deas, E., et al. (2003).Degradation of DIAP1 by the N-end rule pathway is essential for regulatingapoptosis. Nat. Cell Biol. 5, 467–473. doi: 10.1038/ncb984

Do, D., Ostersen, T., Strathe, A., Mark, T., Jensen, J., and Kadarmideen, H. (2014).Genome-wide association and systems genetic analyses of residual feed intake,daily feed consumption, backfat and weight gain in pigs. BMC Genet. 15:27. doi:10.1186/1471-2156-15-27

Do, D. N., Strathe, A. B., Ostersen, T., Jensen, J., Mark, T., and Kadarmideen, H.N. (2013). Genome-wide association study reveals genetic architecture of eat-ing behavior in pigs and its implications for humans obesity by comparativemapping. PLoS ONE 8:e71509. doi: 10.1371/journal.pone.0071509

Fall, T., and Ingelsson, E. (2014). Genome-wide association studies of obe-sity and metabolic syndrome. Mol. Cell. Endocrinol. 382, 740–757. doi:10.1016/j.mce.2012.08.018

Flint-Garcia, S. A., Thornsberry, J. M., and Buckler, E. S. 4th. (2003). Structureof linkage disequilibrium in plants. Annu. Rev. Plant Biol. 54, 357–374. doi:10.1146/annurev.arplant.54.031902.134907

Gilmour, A. R., Gogel, B. J., Cullis, B. R., and Thompson, R. (2009). ASReml UserGuide Release 3.0. Hemel Hempstead: VSN International.

Glas, R., Sauter, N. S., Schulthess, F. T., Shu, L., Oberholzer, J., and Maedler, K.(2009). Purinergic P2X7 receptors regulate secretion of interleukin-1 receptorantagonist and beta cell function and survival. Diabetologia 52, 1579–1588. doi:10.1007/s00125-009-1349-0

Groenen, M. A., Archibald, A. L., Uenishi, H., Tuggle, C. K., Takeuchi,Y., Rothschild, M. F., et al. (2012). Analyses of pig genomes provideinsight into porcine demography and evolution. Nature 491, 393–398. doi:10.1038/nature11622

Haider, S., Ballester, B., Smedley, D., Zhang, J., Rice, P., and Kasprzyk, A. (2009).BioMart Central Portal—unified access to biological data. Nucleic Acids Res. 37,W23–W27. doi: 10.1093/nar/gkp265

Heid, I. M., Jackson, A. U., Randall, J. C., Winkler, T. W., Qi, L., Steinthorsdottir,V., et al. (2010). Meta-analysis identifies 13 new loci associated with waist-hipratio and reveals sexual dimorphism in the genetic basis of fat distribution. Nat.Genet. 42, 949–960. doi: 10.1038/ng.685

Horvath, S. (2011). Weighted Network Analysis - Applications in Genomics andSystems Biology. Los Angeles, CA: Springer. doi: 10.1007/978-1-4419-8819-5

Hou, L., Kongsted, A. H., Ghoreishi, S. M., Takhtsabzy, T. K., Friedrichsen, M.,Hellgren, L. I., et al. (2013). Pre- and early-postnatal nutrition modify geneand protein expressions of muscle energy metabolism markers and phospho-lipid fatty acid composition in a muscle type specific manner in sheep. PLoSONE 8:e65452. doi: 10.1371/journal.pone.0065452

Huttner, W. B., and Schmidt, A. (2000). Lipids, lipid modification and lipid–protein interaction in membrane budding and fission—insights from the rolesof endophilin A1 and synaptophysin in synaptic vesicle endocytosis. Curr. Opin.Neurobiol. 10, 543–551. doi: 10.1016/S0959-4388(00)00126-4

Joshi, A., De Smet, R., Marchal, K., Van De Peer, Y., and Michoel, T. (2009). Modulenetworks revisited: computational assessment and prioritization of model pre-dictions. Bioinformatics 25, 490–496. doi: 10.1093/bioinformatics/btn658

Kadarmideen, H., Von Rohr, P., and Janss, L. (2006). From genetical genomics tosystems genetics: potential applications in quantitative genomics and animalbreeding. Mamm. Genome 17, 548–564. doi: 10.1007/s00335-005-0169-x

Kadarmideen, H. N. (in press). Genomics to systems biology in animal andveterinary sciences: progress, lessons and opportunities. Livest. Sci. doi:10.1016/j.livsci.2014.04.028

Kadarmideen, H. N., Watson-Haigh, N. S., and Andronicos, N. M. (2011).Systems biology of ovine intestinal parasite resistance: disease gene modules andbiomarkers. Mol. Biosyst. 7, 235–246. doi: 10.1039/c0mb00190b

Kaminski, M. M., Sauer, S. W., Kaminski, M., Opp, S., Ruppert, T., Grigaravicius, P.,et al. (2012). T cell activation is driven by an ADP-dependent glucokinase link-ing enhanced glycolysis with mitochondrial reactive oxygen species generation.Cell Rep. 2, 1300–1315. doi: 10.1016/j.celrep.2012.10.009

Keller, L., Xu, W., Wang, H.-X., Winblad, B., Fratiglioni, L., and Graff, C. (2011).The obesity related gene, FTO, interacts with APOE, and is associated withAlzheimer’s disease risk: a prospective cohort study. J. Alzheimer’s Dis. 23,461–469. doi: 10.3233/JAD-2010-101068

Kogelman, L., and Kadarmideen, H. (2014). Weighted Interaction SNP Hub(WISH) network method for building genetic networks for complex diseasesand traits using whole genome genotype data. BMC Syst. Biol. 8:S5. doi:10.1186/1752-0509-8-S2-S5

Kogelman, L. J. A., Kadarmideen, H. N., Mark, T., Karlskov-Mortensen, P., Bruun,C. S., Cirera, S., et al. (2013). An F2 pig resource population as a model forgenetic studies of obesity and obesity-related diseases in humans: design andgenetic parameters. Front. Genet. 4:29. doi: 10.3389/fgene.2013.00029

Kools, P., Van Imschoot, G., and Van Roy, F. (2000). Characterization of three novelhuman cadherin genes (CDH7, CDH19, and CDH20) clustered on chromo-some 18q22–q23 and with high homology to chicken cadherin-7. Genomics 68,283–295. doi: 10.1006/geno.2000.6305

Kwon, Y. T., Xia, Z., Davydov, I. V., Lecker, S. H., and Varshavsky, A. (2001).Construction and analysis of mouse strains lacking the ubiquitin ligaseUBR1 (E3α) of the N-end rule pathway. Mol. Cell. Biol. 21, 8007–8021. doi:10.1128/MCB.21.23.8007-8021.2001

Langberg, K. A., Ma, L., Sharma, N. K., Hanis, C. L., Elbein, S. C., Hasstedt, S. J.,et al. (2012). Single nucleotide polymorphisms in JAZF1 and BCL11A gene arenominally associated with type 2 diabetes in African-American families fromthe GENNID study. J. Hum. Genet. 57, 57–61. doi: 10.1038/jhg.2011.133

Langfelder, P., Zhang, B., and Horvath, S. (2008). Defining clusters from a hier-archical cluster tree: the dynamic tree cut package for R. Bioinformatics 24,719–720. doi: 10.1093/bioinformatics/btm563

Ledent, C., Vaugeois, J.-M., Schiffmann, S. N., Pedrazzini, T., Yacoubi, M. E.,Vanderhaeghen, J.-J., et al. (1997). Aggressiveness, hypoalgesia and high bloodpressure in mice lacking the adenosine A2a receptor. Nature 388, 674–678. doi:10.1038/41771

Matkovich, S. J., Booven, D. J. V., Cappola, T. P., and Dorn, G. W. (2010).Association of an Intronic, but not any exonic, FRMD4B sequence vari-ant and heart failure. Clin. Transl. Sci. 3, 134–139. doi: 10.1111/j.1752-8062.2010.00220.x

Melville, S., and Melville, M. S. (2012). Package ‘NCBI2R’. Available online at:http://cran.r-project.org/web/packages/NCBI2R/NCBI2R.pdf

Meuwissen, T., and Goddard, M. (2010). The use of family relationships and linkagedisequilibrium to impute phase and missing genotypes in up to whole-genomesequence density genotypic data. Genetics 185, 1441–1449. doi: 10.1534/genet-ics.110.113936

Meuwissen, T. H. E., Karlsen, A., Lien, S., Olsaker, I., and Goddard, M. E. (2002).Fine mapping of a quantitative trait locus for twinning rate using combinedlinkage and linkage disequilibrium mapping. Genetics 161, 373–379.

Miller, S. A., Dykes, D. D., and Polesky, H. F. (1988). A simple salting out procedurefor extracting DNA from human nucleated cells. Nucleic Acids Res. 16:1215. doi:10.1093/nar/16.3.1215

Narishima, R., Yamasaki, M., Hasegawa, S., and Fukui, T. (2009). Genetic obesityaffects neural ketone body utilization in the rat brain. Obesity 17, 611–615. doi:10.1038/oby.2008.566

Ota, T. (2013). Chemokine systems link obesity to insulin resistance. DiabetesMetab. J. 37, 165–172. doi: 10.4093/dmj.2013.37.3.165

Qi, L., and Cho, Y. A. (2008). Gene-environment interaction and obesity. Nutr. Rev.66, 684–694. doi: 10.1111/j.1753-4887.2008.00128.x

Segal, E., Shapira, M., Regev, A., Pe’er, D., Botstein, D., Koller, D., et al.(2003). Module networks: identifying regulatory modules and their condition-specific regulators from gene expression data. Nat. Genet. 34, 166–176. doi:10.1038/ng1165

Simonis-Bik, A. M., Nijpels, G., Van Haeften, T. W., Houwing-Duistermaat, J. J.,Boomsma, D. I., Reiling, E., et al. (2010). Gene variants in the novel type 2diabetes loci CDC123/CAMK1D, THADA, ADAMTS9, BCL11A, and MTNR1Baffect different aspects of pancreatic β-cell function. Diabetes 59, 293–301. doi:10.2337/db09-1048

Skaper, S. D., Debetto, P., and Giusti, P. (2010). The P2X7 purinergic receptor: fromphysiology to neurological disorders. FASEB J. 24, 337–345. doi: 10.1096/fj.09-138883

Speliotes, E. K., Willer, C. J., Berndt, S. I., Monda, K. L., Thorleifsson, G., Jackson,A. U., et al. (2010). Association analyses of 249,796 individuals reveal 18 new lociassociated with body mass index. Nat. Genet. 42, 937–948. doi: 10.1038/ng.686


http://cran.r-project.org/web/packages/NCBI2R/NCBI2R.pdf





Spurlock, M. E., and Gabler, N. K. (2008). The development of porcine models ofobesity and the metabolic syndrome. J. Nutr. 138, 397–402.

Stengel, A., Hofmann, T., Goebel-Stengel, M., Lembke, V., Ahnis, A., Elbelt, U., et al.(2013). Ghrelin and NUCB2/nesfatin-1 are expressed in the same gastric celland differentially correlated with body mass index in obese subjects. Histochem.Cell Biol. 139, 909–918. doi: 10.1007/s00418-013-1087-8

Storch, J., and Xu, Z. (2009). Niemann–Pick C2 (NPC2) and intracel-lular cholesterol trafficking. Biochim. Biophys. Acta 1791, 671–678. doi:10.1016/j.bbalip.2009.02.001

Trombetta, M., Bonetti, S., Boselli, M., Turrini, F., Malerba, G., Trabetti, E., et al.(2012). CACNA1E variants affect beta cell function in patients with newlydiagnosed type 2 diabetes. The verona newly diagnosed type 2 diabetes study(VNDS) 3. PLoS ONE 7:e32755. doi: 10.1371/journal.pone.0032755

Uyeda, K., Furuya, E., Richards, C. S., and Yokoyama, M. (1982). Fructose-2,6-P2, chemistry and biological function. Mol. Cell. Biochem. 48, 97–120. doi:10.1007/BF00227610

van de Laar, F. A. (2008). Alpha-glucosidase inhibitors in the early treatment of type2 diabetes. Vasc. Health Risk Manag. 4, 1189–1195.

Veroneze, R., Lopes, P. S., Guimarães, S. E. F., Silva, F. F., Lopes, M. S.,Harlizius, B., et al. (2013). Linkage disequilibrium and haplotype block struc-ture in six commercial pig lines. J. Anim. Sci. 91, 3493–3501. doi: 10.2527/jas.2012-6052

Wang, C., and Guo, F. (2013). Branched chain amino acids and metabolic regula-tion. Chin. Sci. Bull. 58, 1228–1235. doi: 10.1007/s11434-013-5681-x

Weiss, K. M., and Clark, A. G. (2002). Linkage disequilibrium and the map-ping of complex human traits. Trends Genet. 18, 19–24. doi: 10.1016/S0168-9525(01)02550-1

Wen, W., Cho, Y. S., Zheng, W., Dorajoo, R., Kato, N., Qi, L., et al. (2012). Meta-analysis identifies common variants associated with body mass index in eastAsians. Nat. Genet. 44, 307–311. doi: 10.1038/ng.1087

Yang, B., Navarro, N., Noguera, J. L., Muñoz, M., Guo, T. F., Yang,K. X., et al. (2011). Building phenotype networks to improve QTLdetection: a comparative analysis of fatty acid and fat traits in pigs.J. Anim. Breed. Genet. 128, 329–343. doi: 10.1111/j.1439-0388.2011.00928.x

Zeggini, E., Scott, L. J., Saxena, R., Voight, B. F., Marchini, J. L., Hu, T., et al. (2008).Meta-analysis of genome-wide association data and large-scale replication iden-tifies additional susceptibility loci for type 2 diabetes. Nat. Genet. 40, 638–645.doi: 10.1038/ng.120

Zheng, Q., and Wang, X.-J. (2008). GOEAST: a web-based software toolkit forgene ontology enrichment analysis. Nucleic Acids Res. 36, W358–W363. doi:10.1093/nar/gkn276

Conflict of Interest Statement: The authors declare that the research was con-ducted in the absence of any commercial or financial relationships that could beconstrued as a potential conflict of interest.

Received: 20 May 2014; paper pending published: 11 June 2014; accepted: 20 June2014; published online: 09 July 2014.Citation: Kogelman LJA, Pant SD, Fredholm M and Kadarmideen HN (2014) Systemsgenetics of obesity in an F2 pig model by genome-wide association, genetic network,and pathway analyses. Front. Genet. 5:214. doi: 10.3389/fgene.2014.00214This article was submitted to Systems Biology, a section of the journal Frontiers inGenetics.Copyright © 2014 Kogelman, Pant, Fredholm and Kadarmideen. This is an open-access article distributed under the terms of the Creative Commons Attribution License(CC BY). The use, distribution or reproduction in other forums is permitted, providedthe original author(s) or licensor are credited and that the original publication in thisjournal is cited, in accordance with accepted academic practice. No use, distribution orreproduction is permitted which does not comply with these terms.


http://dx.doi.org/10.3389/fgene.2014.00214



http://creativecommons.org/licenses/by/3.0/







Documents

Systems genetics of obesity in an F2 pig model by genome-wide association, genetic network, and pathway analyses